Autodesk Stingray and the future

[Jon Berntsen]

Now as the Stingray engine starts to mature, how are you guys looking at it as a competitor to Unreal Engine?

 

For me, Stingray starts to unveil itself with a lot of opportunities for us who already work with max and vray. Looks like a platform for the future to me. Obviously the biggest issue with realtime engines is that you still need engines like vray or corona to do good high resolution still images. So you're bound to two platforms. So if you want to make add/transform your project into a realtime one, Stingray should be right on the spot. What do you think? Will Stingray still be around 5 years from now?

[Francisco Penaloza]

I been using Stingray for a year now, and I am pretty happy with it so far, I think I have more flexibility to create stuff inside Stingray that I did in Unreal or Unity.

Now Unreal is a way more mature software, no questions you can do much more stuff than is not possible on Stingray yet, but since I am not a programmer I found stingray easy to learn and quicker to do simple stuff.

 

Right now what really make a difference is the large community that Unreal has, you can find almost anything pre-made, it is like comparing VRay-Max with Cinema-Redshift.

 

I work in an Architectural Company so for other users my opinion won't apply of course. but I believe if Autodesk don't mess up, Stingray could become the go to tool to do Real-time ArchViz or VR ArchViz. Exporting REVIT model in to Stingray through 3D Max it pretty straight forward.

What I am afraid is that Stingray get shrinked down to a one click solution for the AEC industry with no options .

 

Regarding comparing Raytracer rendering with real time rendering, I believe that our industry will change a lot in the coming years.

Considering the quality that you can get out of Unreal vs how long it take you to prep and render an animation in VRay, I think those times are changing rapidly in favor of real time engines, it is just crazy.

[John Dollus]

toolsets purchased by AutoDesk don't have a very good track record. They tend to loot the code, integrate what they want/need for their flagship products and then kill off the donor. As such, I would not even consider Stingray until after a proven 3 year dev record under their umbrella.

Other solutions (like vray and corona) are rapidly closing in on solutions that will come very close or even surpass what Stingray has to offer the Arch sector.

[Jon Berntsen]

toolsets purchased by AutoDesk don't have a very good track record. They tend to loot the code, integrate what they want/need for their flagship products and then kill off the donor. As such, I would not even consider Stingray until after a proven 3 year dev record under their umbrella.

Other solutions (like vray and corona) are rapidly closing in on solutions that will come very close or even surpass what Stingray has to offer the Arch sector.

 

Interesting. Do you really think Vray is adding functionality to become a game engine like Stingray or Unreal Engine? And as per now, it's pretty fantastic hard to get even good dynamic atmospherics i Vray, like weather and skies, if at all. I haven't seen any sign that they're even adding this? Things are happening rapidly now, and I can't sit and wait.

 

Can you refer to which functions Vray has added to take this step?

Edited January 13, 2017 by chroma

[Francisco Penaloza]

Maybe he is talking about of speed/workflow optimization instead of real-time development.

I am pretty sure that they could jump on that too, there are pretty cool research already on real-time GI, the results are very promising, That's why I think in the close future real-time engines will catch up with the raytracer flexibility and not the other way around.

either way we win

 

 

I know Autodesk had done strange stuff in the past but at the same time I need to produce now and not in 3 years, considering the time that take to learn Stingray I would not mind if it goes down on 3 years, by then everything could be easily translate to Unreal or Unity. besides you should never put all your egg on a single engine, at my office we use, V-Ray and Corona, some old project still are worked on Mental Ray, and I am still doing some stuff in Unreal at the same time I do production on Stingray. they are just tool, use them as you need them.

[Jon Berntsen]

I appreciate your answer.

What do you mean with this?

 

I know Autodesk had done strange stuff in the past but at the same time I need to produce now and not in 3 years, considering the time that take to learn Stingray I would not mind if it goes down on 3 years, by then everything could be easily translate to Unreal or Unity. ...

 

Do you mean that in 3 years, you believe that the advantage Max Stingray with Vray support is being faded out, so Unreal Engine will have the a similar support to make things easier?

[Francisco Penaloza]

I appreciate your answer.

What do you mean with this?

 

 

 

Do you mean that in 3 years, you believe that the advantage Max Stingray with Vray support is being faded out, so Unreal Engine will have the a similar support to make things easier?

 

No, My point was that John mentioned that in 3 year (or whatever number) Autodesk could dump Stingray and that's why he didn't want to invest time until Autodesk prove support for this software. I truly understand his position, I am very skeptical too. But in my case I need it to move a head with VR presentations and using Unreal was a very slow learning curve, that's why I give a try to Stingray. It worked out pretty fast actually.

I am not saying it is a better software, and I also don't know if it will last forever or just a few years. What I am saying is if you are curious of try any Game engine, you should try Stingray it is a pretty solid option right now.

 

My other comment was more about, whatever you learn while using on Stingray can be applied in any other Game engine such Unreal or Unity.

[philippelamoureux]

The best of both worlds will be Octane in unity (coming in 2017) and Octane in ue4 plugin, with no ETA at this time.

[Scott Schroeder]

Stingray is geared towards a completely separate need in architecture. It's for rapid development of a space, yet you need to be able to walk around it to really grasp the design. It's similar to IrisVR and their Revit tools. Quality is only somewhat important but speed and how quick it takes to get from software A to software B are the major driving factors.

 

UE4/Unity only really come into place once the design is 101% complete. Stingray can fill in that gap during the years of design development.

 

Stingray is a great tool for early design development. Though, I too am weary about Autodesk's track record with their softwares talking to each other. I still have the great FBX link scale f-up from 2016 still in my head.

 

I'm not holding my breath for Octane anytime soon. Yes boss, can I get approved on spending 20k on graphics cards?

[Jon Berntsen]

Stingray is geared towards a completely separate need in architecture. It's for rapid development of a space, yet you need to be able to walk around it to really grasp the design. It's similar to IrisVR and their Revit tools. Quality is only somewhat important but speed and how quick it takes to get from software A to software B are the major driving factors.

 

UE4/Unity only really come into place once the design is 101% complete. Stingray can fill in that gap during the years of design development.

 

Stingray is a great tool for early design development. Though, I too am weary about Autodesk's track record with their softwares talking to each other. I still have the great FBX link scale f-up from 2016 still in my head.

 

I'm not holding my breath for Octane anytime soon. Yes boss, can I get approved on spending 20k on graphics cards?

 

So if you were supposed to jump on the real-time train for customer projects, your weapon of choice would be?

 

I hear talk about unreal engine being able to spit out basically any image or animation from it, in very competitive quality, if you just have the scene set up. Then it could outrun our existing max/vray workflow.

 

However, setting up completely finished buildings for real-time is one thing. But then we would need to employ a huge library, and it would be stupid to not use the same platform for customer projects as well. Or said in another manner, it would be stupid to maintain both vray and unreal props-libraries. So this is where I thought Stingray could be the answer, since it supports vray materials. But these game engines doesn't like the typical hipoly max-furnitures anyway, right? A typical evermotion chair would be of no use? If that's how it is, then the other option is also ruled out, which is setting up neat hi end hires stuff in max/vray, and then transfer the same assets to Stingray to further build the real-time functionality.

 

I guess the real question here is: When should we start using game engines instead of the slow raytracers like vray? Imagine being able to set up one house in Unreal on stingray and you've got as many animations from wherever you want, as many stills as you want, and you can also export 360 stills for webvr use. Isn't this the future?

 

Many question marks in there, but that really reflects me on if I'm going to start developing visualization projects on game engines (and which one), 100% all in. The prize can be really nice. My boss has given me one week to find out if we should start developing on unreal, stingray og whatever we should do to be able to make both nice images and fully set up interiors in real-time engine. How can those tasks possibly be merged?

 

Sky/cloud system is also a big deal. Vray is not even close to having this. And renderings take 4-5 hours or more per still, minimum. Movie play time, 5 minutes... Try to get that out of vray. With atmospheric effects, clouds, burning flames in the fireplace, moving grass and trees, etc etc

Edited January 13, 2017 by chroma

[Francisco Penaloza]

Well I don't think anybody has a straight answer to those questions Honestly, mostly because it all depend of what type of work your do, this mean, are you a niche interior/ furniture viz artist a la Bertrand? or you do large project and developments a la Neoscape or Crystal, or you are Architecture in house viz guy that has to spit "exiting images' with almost no time?

 

Without thinking in details I think Unreal is the software that seems more rounded and mature today, there are tonz of great example showcasing how far it can go, with know how and time. But if you don't have one of those, then things fall a part.

 

I am a in house viz guy for now. and I would love to do a nice Unreal Scandinavian house but my reality are, very messy Sketchups, Rhino, Bonsai, REVIT files coming from Architects and Engineers. They all are "ready to render" and they all need it now... within budget of course.

 

Because I am in that environment Stingray work fine, I can spit really crappy looking real-time or VR presentations, and everybody is happy. if I have time I can make it look good . If I want them to look like Koolas Barcelona Pavilion I would run out of time and the building I receive are far from ready for something like that.

 

If you are a medium Arch Viz only company, then maybe Unreal can be a better choice but you'll need a team, some one who could prep all your exiting libraries from VRay and make them Unreal friendly(reduce poly count, LOD, Normal Maps, Shaders). you'll also need a Math headed guy who can take care of Blueprints and hopefully C++.

 

If you have a good team in theory you could compete on production time with a VRay presentation, with all the extras as you mentioned that you could get such animations, still and VR or Web or Ipad and what not. even if it take you a little longer you could get a lot more, but a single man can't do that within schedule.

 

This guy from UE4Arch does pretty good work, but he said it take a month or so to produce a medium size project. A loft, a simple house in the Forrest, and Apartment interior of 900 SFT. Forget about a large high rise in the middle of Manhattan; you'll need all Crystal CG sweat shop, I mean Team to help you in that one.

 

Honestly I think for companies dedicated only to ArchViz, that mean final end product, it is hard to jump in to VR and real-time production now. Because it is a big investment, for in House Architectural firms or engineering or similar, we could get away with good enough most of the time and from time to time we could spend more time to get something great to save our artistic souls.

 

For in House I would totally recommend Stingray, I have a full REVIT Hospital 5 Stories Xref in to 3D Max and exported to Stingray for real-time fly around in a couple of days, shoot if they don't care about quality wise it can be in hours. But it is not pretty and is choppy but hey is real-time.

[philippelamoureux]

What I can say concerning the design process, when you want a quick view and quality isn't important, unreal can be pretty damn fast. Export/import fbx as a combined mesh takes seconds. You can use dynamic lighting like lumion does, it's basic but it gives you an overview. It's all VR ready. Use the official VR Template, import fbx, scale it right, bam you are in VR! Now even the editor itself has a VR mode if you want. You could be Inside your building and place furnitures, apply materials, scale things, etc while in vr. Gimmicky but still cool.

 

 

What takes ''ages'' in unreal is the high end stuff, like ue4arch. It's the marketing imagery.

 

 

A feature of unreal that is maybe not well known is that once you import textures, objects, etc. in the editor and you make a change to any source file in another software, like photoshop for example, Unreal automatically look for those changes and update them in the ue4 editor.

 

 

Eventually we'll have real-time G.I and I think the pioneers are going to be game engines. So, what you learn today will serves you in the future!

Edited January 14, 2017 by philippelamoureux

[Jon Berntsen]

Thanks for input. I just read the road map for vray 3.5, and it says they will implement a support for rendering in vrayRT directly to vive and oculus. Wow! But I wonder how much interactivity that can be made out of it, like changing kitchen sets or turning on and off lights on the fly, or changing materials. It should all be depending on if it's possible to programming it? Sounds exciting to me. Although I need to test vrayRT. How real-time can it be, I believe vrayRT is far far away from a game engine.

[Jon Berntsen]

Thanks for input. I just read the road map for vray 3.5, and it says they will implement a support for rendering in vrayRT directly to vive and oculus. Wow! But I wonder how much interactivity that can be made out of it, like changing kitchen sets or turning on and off lights on the fly, or changing materials. It should all be depending on if it's possible to programming it? Sounds exciting to me. Although I need to test vrayRT. How real-time can it be, I believe vrayRT is far far away from a game engine.

 

Installed vray 3.5 beta. Well, all I have to say that this is not going to be a good solution for realtime VR.

[philippelamoureux]

Even if their interactive renderer can sometimes almost seems like real-time, vrayRT, corona interactive are nothing like a game engine. It's just a renderer. Unity, Unreal, etc are much more than a renderer.

 

I suppose their rendering directly for vive/rift is just the interactive renderer but in stereoscopic pano? I am wrong? The noise takes forever to clear?

[Jon Berntsen]

Even if their interactive renderer can sometimes almost seems like real-time, vrayRT, corona interactive are nothing like a game engine. It's just a renderer. Unity, Unreal, etc are much more than a renderer.

 

I suppose their rendering directly for vive/rift is just the interactive renderer but in stereoscopic pano? I am wrong? The noise takes forever to clear?

 

Yes so if you're moving 1 cm, it has to render and clear up again. Forever, well that's a broad term, but at least it's not performancing 60 f/s.

[philippelamoureux]

The difference is also that with vray/corona stereo pano, it's just a spheric image wrapped around you and you are always at the center (camera view). Each time you move it has to re-render the whole 360 image. In game engines you are in the actual scene. You render the scene once and then you can do whatever you want in it. It's not bound to the camera perspective. If that make sense.

 

If you have a rift/vive try the default vr template from ue4, it's pretty cool.

[Jon Berntsen]

The difference is also that with vray/corona stereo pano, it's just a spheric image wrapped around you and you are always at the center (camera view). Each time you move it has to re-render the whole 360 image. In game engines you are in the actual scene. You render the scene once and then you can do whatever you want in it. It's not bound to the camera perspective. If that make sense.

 

If you have a rift/vive try the default vr template from ue4, it's pretty cool.

 

I'm testing in UE4 right now. Just with plain simple building though, and not VR. Have opted to buy a Vive, so hopefully I will get to that very soon! Seems promising, although a steep learning curve.

[Scott Schroeder]

What I can say concerning the design process, when you want a quick view and quality isn't important, unreal can be pretty damn fast.

 

In my 16 years of doing this, I can say for certain that UE4 isn't fast enough for most places, especially in-house. It's the design process and changes that are going to get you.

 

VR and real time are important for the end stage needs of architectural visualization. They have no place in the front design stages right now, unless you have one-click solutions like Stingray and say IrisVR. We're testing Enscape at work and I'm blown away at how well you can move around in our stadium Revit models in near real time.

[philippelamoureux]

ok I understand how you guys need a very very fast solution, I guess the fact that materials and lights don't have to be redone a second time is a big factor too. I've asked the guys at irisVR what was the tech behind the 1-click vr solution and this is what they told me last summer.

 

Hi Philippe,

 

With Prospect, you can covert your Revit, Sketchup, and OBJ files (with more file types coming this year), into fully navigable (like Unity and Unreal) experiences for the Vive.

The immersive experience is powered by Unity through our own proprietary file type and our own technology built into it. For users, Unity itself is never exposed. I hope these helps.

Our latest version was released today so you can download it and see for yourself.

 

The one caveat I would offer is that we currently do not support Revit 2017, but will soon.

Looking forward to hearing from you.

 

Best,

George

 

George Valdes -

IrisVR

[Jon Berntsen]

Scott, makes sense. Never heard about Escape, but probably not the only product I haven't heard about these days. I have needs that covers both typical late stage of in-house from our architects, plus sales visualization work for external customers. I tested unreal today, and I was able to import my entire interior that I've set up in max, as a 500 mb fbx. And it worked. Also exteriors worked, with custom modeled cladding as as well. No textures or shaders, so this has to be made, catalogued and structured in ue4, but once that is done, it's mostly straightforward putting on any texture.

 

Using ue4 for hi end stills images sounds like a no go to me, after testing.

[Scott Schroeder]

In case you haven't seen this video.

 

We're not yet at that point for full VR in hyper real quality. Only a small few projects have the timeline to support this currently. Heck, even Brick's latest VR app is still based on superb quality static pano's in VR.

 

VR may start to trend towards mobile, so it may be some time before they really solve the full space tracking that mobile based VR lacks. Why mobile? Well, data is starting to come out that the full VR headsets like the Vive have fallen flat in sales on the consumer side. This is in part to the expensive hardware needed to power VR scenes. That doesn't work well for us either, because I don't feel like looking like Hodor lugging around my Bran computer on my back just to show VR for a client at their office or whatnot. It's far easier for me to whip out the phone, into the headset, into VR for the client. It's also far easier to pitch to a client to buy a $99 headset versus a $2-3k computer on top of a $600 headset.

 

Of course, none of this matters until they can solve the sickness issues that many people experience if they move too much in VR.

 

https://www.nytimes.com/2017/01/08/business/sticker-shock-and-maybe-nausea-hamper-sales-of-virtual-reality-gear.html

[Rich O]

toolsets purchased by AutoDesk don't have a very good track record. They tend to loot the code, integrate what they want/need for their flagship products and then kill off the donor. As such, I would not even consider Stingray until after a proven 3 year dev record under their umbrella.

Other solutions (like vray and corona) are rapidly closing in on solutions that will come very close or even surpass what Stingray has to offer the Arch sector.

 

What do V-ray have in the pipeline? I don't *recall* seeing Vlado or anyone mention anything in that regard, at least on the forums

[James Vella]

thanks for sharing Scott! Its always interesting to see other studios workflows

[Jon Berntsen]

What do V-ray have in the pipeline? I don't *recall* seeing Vlado or anyone mention anything in that regard, at least on the forums

 

Yeah, seems like nothing much. However, since realtime engines are limited in its output resolution, they can't compete with raytracers like vray anyway. Our customers needs images in the size of 6000px +. But the realtime engines will be there in a couple of year, I guess. The problem is also that geometry needs to be optimized anyway, so using the same scene won't be possible no matter what. So, really, when realtime engines are "up to date" on the resolution needs, there's no going back.